2016 Volume 63 Issue 7 Pages 573-578
A newly developed Ni-free pre-alloyed steel powder, Fe-0.5 %Cr-0.2 %Mn-0.2 %Mo (JIP® 5CRA) shows a rotating bending fatigue limit of 510 MPa after compacting, sintering and case-hardening process. This is approximately 40 % higher than that of a case-hardened sintered steel based on a conventional diffusion bonded steel powder, Fe-4 %Ni-1.5 %Cu-0.5 %Mo (4Ni). This is attributed to the higher compressive residual stress of the 5CRA based material. In spite of the higher fatigue limit, its tensile strength is lower than that of the 4Ni based material. The difference in the microstructural transformation behavior is a possible cause of this reverse phenomenon.
PM technology is featured with its capability of near net shape forming, and with this feature, Fe-based sintered materials have shown significant advantages for producing components such as automotive parts with complex geometries1). According to demands for improvement of vehicle fuel efficiency, high strength sintered materials are required for downsizing and weight reduction of components.
A diffusion bonded alloyed powder with a composition of Fe-4 %Ni-1.5 %Cu-0.5 %Mo (4Ni) has been one of the most widely applied alloyed powders for such high strength applications2,3). However, according to the demands for Ni reduction, Cr and Mn have been investigated as the alternative alloying elements for their hardenabilities higher than that of Ni. One problem is that Cr and Mn are tend to be oxidized during the sintering and/or heat-treatment process because of their strong affinity to oxygen, and their oxides would give negative effects on mechanical properties. Nevertheless, because of the technological advances in sintering and heat treatment processes, it has become possible to use Cr and Mn effectively as strengthening elements4–8). Maetani et al.9) have introduced a Cr-prealloyed powder JIP® 5CRA with a Ni-free composition of Fe-0.5 %Cr-0.2 %Mn-0.2 %Mo, and shown an as-sintered tensile strength equivalent to that of the conventional 4Ni diffusion bonded alloyed powder.
In this study, a case-hardening heat-treatment was applied to further improve the mechanical properties of the 5CRA based sintered material. The case-hardened material showed a fatigue limit of 510 MPa and a tensile strength of 1250 MPa. The fatigue limit ratio is also significantly higher than that of the 4Ni based material. This is attributed to the difference in the microstructure, and the eventual difference in the microstructural transformation behavior during the strain process.
Two steel powders, a prealloyed Fe-Cr-Mn-Mo powder (JIP® 5CRA) and a partially diffused Fe-Ni-Cu-Mo powder (JIP® Sigmaloy® 415S) were provided as the raw materials of this study. The compositions of the powders are shown in Table 1. Total amounts of alloying elements are 0.9 % for 5CRA and 6 % for 415S. These powders were mixed with natural graphite powder (mean particle size: 5 μm) and lubricants into the compositions shown in Table 2. These powder mixtures were compacted into two types of bar-shaped specimens with dimensions of 15 × 15 × 80 mm (bar A) and 10 × 10 × 55 mm (bar B). The green density was adjusted to 7.15 Mg/m3. The bar-shaped specimens were sintered at 1200 °C for 150 min in 90 vol% N2 + 10 vol% H2 atmosphere. Table 3 shows the density and carbon content after sintering. The sintered densities of the both materials were equivalent with each other, however, the carbon loss of 0.5Cr during sintering was larger than that of 4Ni because of higher oxygen content in the 5CRA.
| Alloyed powder | Alloying elements (mass%) | |||||
|---|---|---|---|---|---|---|
| Ni | Cu | Cr | Mn | Mo | sum | |
| JIP® 5CRA | — | — | 0.51) | 0.21) | 0.21) | 0.9 |
| JIP® 415S | 4.02) | 1.52) | — | — | 0.52) | 6.0 |
| Sample | Alloyed powder | Graphite (mass%) | Lubricant (mass%) |
|---|---|---|---|
| 0.5Cr | JIP® 5CRA | 0.8 | 0.5 (Wax) |
| 4Ni | JIP® 415S | 0.4 | 0.8 (ZnSt) |
| Sample | As sintered | |
|---|---|---|
| Sintered Density (Mg/m3) | C content (mass%) | |
| 0.5Cr | 7.24 | 0.43 |
| 4Ni | 7.25 | 0.31 |
Fatigue test pieces with a diameter of 8.0 mm were machined from the as-sintered bar A specimens. Tensile test pieces with a diameter of 5.0 mm were also machined from the as-sintered bar B specimens. Heat treatment (carburizing: 870 °C × 60 min, carbon potential: 0.8 mass%, oil quenching: 60 °C, tempering: 180 °C × 60 min in air) was applied to these machined test pieces.
Fatigue tests were carried out under the condition of rotation speed of 3000 min−1 and stress ratio R = −1. The fatigue limit was determined as the maximum endurance stress at 107 cycles. Tensile tests were carried out at a tensile speed of 5 mm/min with the repetition number 5 for each material. Yield strengths in this study were determined as the stress at the point where the stress-strain curve departs from the straight line in the elastic region.
The sintered bar B specimens were also case-hardened, and their hardness and microstructure were evaluated on the cross section at the long side center. The hardness depth profile was measured with a micro-Vickers hardness meter with a load of 0.49 N. Optical and scanning electron microscopic observations were carried out with the nital-etched cross section. Pore size and circularity were also evaluated from the optical micrographs of the non-etched cross sections with an image analysis software (Image J). The Feret diameters, the perimeters and the areas of all the pores included in the observed area (more than 750 pores) were measured.
The amount of retained austenite was measured on the side surface of the case-hardened bar B specimens by X-ray diffraction. The residual stresses along the axis of the fatigue test pieces were measured on the surface of the test pieces by X-ray micro-stress meter.
The S-N curves obtained from the rotating bending fatigue tests are shown in Fig. 1. The fatigue limit of 0.5Cr is 510 MPa, about 30 % higher than that of 4Ni (400 MPa) even with its lower amount of alloying elements. Fig. 2 shows the yield and tensile strengths. The error bars show the standard deviations (±1σ) over the repetition number of 5. The yield strength of 0.5Cr is 1120 MPa, about 40 % higher than that of 4Ni (822 MPa). The tensile strength of 0.5Cr, however, is 1250 MPa, about 10 % lower than that of 4Ni (1370 MPa). From these results, the fatigue strength ratios are calculated as 0.41 for 0.5Cr and 0.29 for 4Ni. More than 40 % higher fatigue strength ratio is obtained with 0.5Cr comparing to 4Ni.
S-N curves of rotating bending fatigue test for case-hardened 0.5Cr and 4Ni.
Yield and tensile strengths of case-hardened 0.5Cr and 4Ni.
Fig. 3 shows the hardness profiles from the edge to the center of the cross section of the case-hardened bar B specimens. For both 0.5Cr and 4Ni, the hardness gradually decreases and levels off at a distance of 2 mm from the edge. The hardness of 0.5Cr is 100HV higher than that of 4Ni over the measured area.
Vickers hardness profile of case-hardened 0.5Cr and 4Ni.
Microstructures of 0.5Cr and 4Ni are shown in Fig. 4. Tempered martensite dominates over the observed area for 0.5Cr, however, there are bainite phases as well in the microstructure of 4Ni. In addition, as shown in Fig. 5, 0.5Cr and 4Ni contain 4.6 vol% and 8.1 vol% austenite, respectively. Consequently, 0.5Cr contains more tempered martensite than 4Ni. The residual stresses of the surface of fatigue test pieces are shown in Fig. 6. Observed stresses are compressive for both materials, and the values are 709 MPa and 661 MPa for 0.5Cr and 4Ni, respectively.
Microstructures of case-hardened 0.5Cr: (a) and 4Ni: (b).
Amounts of retained austenite of case-hardened 0.5Cr and 4Ni.
Residual stresses of case-hardened 0.5Cr and 4Ni.
The mechanical properties of the two materials are summarized in Table 4. As mentioned above, 0.5Cr shows that a fatigue limit is higher than that of 4Ni. The sintered densities of both materials are almost equivalent as shown in Table 3. It means the total volume fractions of the pores in each material are equivalent. If we regard the Feret diameter as the pore size, the number fraction distribution is calculated from the image analysis results as shown in Fig. 7 which says the pore size distributions of both materials are also equivalent to each other. The circularity of each pore is defined as
| Sample Alloyed powder |
0.5Cr JIP® 5CRA |
4Ni JIP® 415S | |
|---|---|---|---|
| Fatigue strength, σW (MPa) | 510 | 400 | |
| Tensile strength, σT (MPa) | 1120 (±126) | 822 (±133) | |
| Yeild strength, σY (MPa) | 1250 (±80) | 1370 (±12) | |
| Microstructure (in appearance) | by OM and SEM | tempered martensite | tempered martensite + bainite |
| Amount of retained austenite (vol.%) | by X-ray diffraction | 4.6 | 8.1 |
| Vickers Hardness (HV0.05) | Surface | 800 | 700 |
| Center core | 650 | 550 | |
| Compressive residual stress, σr (MPa) | 709 | 661 | |
| Pore size distribution | equivalent | ||
| Circularity of pores | 0.85 | 0.85 | |
( ): standard deviation
Pore size distributions and circularities of case-hardened 0.5Cr and 4Ni.
| (1) |
where A and P are the area and perimeter of the pore, respectively, and evaluated from the image analysis results. The average values of circularities over the observed pores were 0.85 for both materials. Consequently, it is assumed that the cause of the differences of the fatigue limits of 0.5Cr and 4Ni is not based on the pore volume fraction, the pore size distribution or circularity.
Matsui et al.10) have proposed a fatigue limit model with yield strength and compressive residual stress for wrought steels. The yield strength (σY) and the residual stress (σRmax) represent the resistivity factor against crack generation (stage I) and crack propagation (stage II), respectively, and they derived a linear relationship for the fatigue limit (σW) as
| (2) |
where K is a constant and σRmax is the maximum value of compressive residual stress in the depth direction from the surface of the specimen. A wrought steel shows a good agreement with this model with the constant K of 0.3848.
If we apply the same model to the materials in this study with parameters in Table 5, where σr was substituted for σRmax, it also shows a good agreement with K value of 0.2752 as shown in Fig. 8. Consequently, the fatigue limits of 0.5Cr and 4Ni, which are sintered steels, are fundamentally based on the same mechanism as wrought steels. However, the slope of the straight line of sintered materials (0.2752) is smaller than that of wrought materials (0.3848). It would be attributed to the pores contained in sintered materials, which can accelerate the crack generation and propagation in fatigue tests to make fatigue limit decrease.
| Sample | σY (MPa) | σRmax(σr) (MPa) | σY + σRmax (MPa) | σW (MPa) |
|---|---|---|---|---|
| 4Ni | 822 (±126) | 661 | 1483 | 400 |
| 0.5Cr | 1120 (±133) | 709 | 1829 | 510 |
( ): standard deviation
Relationship among fatigue limit, yield strength, and residual stress of case-hardened 0.5Cr and 4Ni.
In the discussion above, we assumed σr at the surface as σRmax of the sintered steels. The compressive residual stress arises from martensite transformation which is accompanied by volume expansion. And the cooling rate is the highest at the surface of the specimens where there is the most amount of martensite leading to the highest compressive residual stress. Therefore, the substitution of σr for σRmax would be a rational assumption in this case. Furthermore, 0.5Cr has a larger compressive residual stress than 4Ni. For the explanation for that, martensite transformation occurs more in 0.5Cr due to its Ni-free composition.
The hardness profile in Fig. 3 can also be explained from the aspect of martensite transformation. The hardness is the highest at the surfaces of the specimens for both materials. As mentioned above, it is due to the most amount of martensite at the surface of the specimens. And 0.5Cr is harder than 4Ni, because martensite transformation occurs more in 0.5Cr and there are austenite and bainite phases more in 4Ni.
4.2 Tensile StrengthAs mentioned above, 0.5Cr has a higher fatigue limit than that of 4Ni, however, it shows a lower tensile strength. It seems against the general expectation that a higher fatigue limit is accompanied by a higher tensile strength11). This reverse phenomenon would be attributed to the higher amount of retained austenite in 4Ni. Retained austenite causes Transformation Induced Plasticity (TRIP) effect to increase tensile strength. According to Furukimi et al.12), tensile strength is determined by volume fraction of martensite transformed from austenite though the strain induced transformation effect.
In order to confirm the strain induced transformation, the retained austenite of the specimens after the tensile test was measured by X-ray diffraction with spot diameter of 1 μm. Polished cross section parallel to the axis of the fractured tensile specimens were applied for this measurement. As shown in Fig. 9, the amounts of retained austenite in the vicinity of the fractured surface are 2.8 vol% and 1.9 vol% for 0.5Cr and 4Ni, respectively. By calculating the difference from the values before the tensile test in Fig. 5, the amount of transformed retained austenite in 4Ni during the tensile test is obtained as 6.2 vol%, which is more than three times larger than that of 0.5Cr (1.8 vol%). Hence, TRIP effect works more in 4Ni leading to its higher tensile strength.
Schematic diagram showing the measured points and amounts of retained austenite in test pieces after tensile test for case-hardened 0.5Cr and 4Ni.
Fig. 10 shows an example of true stress-true strain curves of 0.5Cr and 4Ni. The Young’s modulus (the slope of the dashed line in Fig. 10) of 0.5Cr is larger than that of 4Ni. All the Young’s moduli were read off from all of the true stress-true strain curves from the repetition number of five for each material. The average values were calculated as 146 GPa and 108 GPa, and the standard deviations were also were calculated as 37 GPa and 20 GPa for 0.5Cr and 4Ni, respectively. There is statistically a significant difference in level of significance 0.1 between the two average values.
Examples of true stress-true strain curves obtained in the tensile test for case-hardened 0.5Cr and 4Ni.
An ultrasonic pulse-echo measurement has given a relationship between Young’s modulus E and sintered density ρ of sintered steels as
| (3) |
where E and ρ are in the units of GPa and Mg/m3 13). The calculated value 159 GPa for 0.5Cr is equivalent to the experimental value of 146 GPa within the range of experimental error. This indicates that the tensile properties of 0.5Cr are mainly dominated with the pore fraction. For 4Ni, however, the experimental value of 108 GPa is significantly smaller than the calculated 160 GPa from eq.(3). This would be also attributed to the larger amount of strain induced transformation in 4Ni. The strain induced transformation is accompanied by volume expansion. Therefore, the stress increase can be more easily released in 4Ni during tensile test, resulting in a stress smaller than that of 0.5Cr at the same strain.
The mechanical properties of the case-hardened low Cr sintered steel (0.5Cr) were compared with those of the conventional 4 %Ni case-hardened sintered steel (4Ni).
(1) Fatigue limits were 510 MPa for 0.5Cr and 400 MPa for 4Ni. Despite the low amount of alloying elements, 0.5Cr has better fatigue properties. It is attributed to its higher compressive residual stress resulting from the higher amount of tempered martensite.
(2) Even though 0.5Cr has more tempered matensite and a higher hardness than 4Ni, it has a lower tensile strength of 1250 MPa than 4Ni (1370 MPa). The lower amount of martensite transformed from retained austenite is a possible cause of the lower tensile strength of 0.5Cr.
